NetNORAD – Facebook TSHOOT tool

Interesting read and video about NetNORAD – a tool used by facebook to detect and isolate network latency/packet drop issues.

https://code.facebook.com/posts/1534350660228025/netnorad-troubleshooting-networks-via-end-to-end-probing/

 

Additional reads:

https://github.com/facebook/fbtracert

https://github.com/facebook/UdpPinger

Accessing HomePC from Anywhere[without traffic passing through a third-party app]

• Create a free account with a Dynamic DNS service like noip.com and configure it in the modem. This step is to link dynamically assigned IP from your service provider to a static name.
• Install any remote sharing app like VNC server, ultraVNC, windows RDP in the HomePC.
• Configure static DHCP allocation in the wireless router such that same local IP is assigned/reserved to HomePC.

Linux DHCP server config for PXE/BMP boot

Dell switches support Bare-metal provisioning(BMP) where a switch can be provisioned to automatically download image and config file from a remote server and boot up with that image file. In a nutshell, the process involves:

IPv6 – Notes

• IPv6: 128 bits. Represented in hexadecimal format as eight 16 bits segments.
• Two rules for IPv6 representation:
  • Leading zeros in 16 bits can be represented as single 0.
  • Single contiguous string of one or more segments are all zeros can be represented as ::

BGP Notification Message

A BGP notification message is sent when an error condition is detected by the BGP process. The TCP connection with that BGP peer is closed after sending the notification message. Usually, the BGP process will log this notification in syslog, which would help us to identify the reason why the BGP session was brought down.

In this post, BGP notification message format along with error codes and subcodes are discussed.

Python Script that SSH to a switch and execute commands for ‘x’ times

Below is a sample python code that can login to a switch via SSH and gather few command outputs and display it on console. SSH sessions are repeated for ‘x’ number of times.

Paramiko SSH module is used in this script. Paramiko is a native Python implementation of SSH. This module can be installed on a linux system using pip. commands are:

#sudo apt-get install python-pip
#sudo pip install paramiko

Code:

Network Troubleshooting Methods

Troubleshooting a network connectivity issue is quite challenging in a complex Data Center Network with live production traffic.

Below are few basic methods to identify a node in the traffic path that drops packets which would cause network connectivity issues. This procedure is in general and not related to any specific vendor as troubleshooting options varies with each vendor.

Let’s consider below simple topology for troubleshooting purpose:

R1(IP:1.1.1.1)— R2 — R3 — R4(IP:2.2.2.2)

Problem statement#1: 100% ping loss between R1 and R4. In this case, ping from R1 to R4’s interface IP fails 100%.

NPG (NPIV Proxy Gateway) Mode Vs Full Fabric Mode in FCoE

In this article, I will be discussing about two different modes that an FCF switch can operate in FCoE topology.

FCF(FCoE Forwarder) is a switch that supports both Ethernet(FCoE) functionality like FIP, DCBx, LAN protocols and FC(Fiber Channel) functionality like fabric login, zoning and name services. FCF  connects the CNA on the server end to any storage array(Native FC), by bridging the gap between Ethernet LAN and storage . Dell S5000 is an example of FCF switch which can be configured in two different modes- NPG or Full Fabric mode.

Port names in FC/FCoE Topology

Following are most common port names used in FC/FCoE Technology:

N_port(Node port)/VN_port:

• Ports on end host(i:e HBA) and storage array(target port) are called as N_port in FC topology.
• Called as VN_port in FCoE terminology.
• HBA’s that handles IP traffic are called as CNA in FCoE terminology.
• Each N-port will have a unique WWPN (64 bit) assigned by vendor. We use this WWPN while configuring Zoning on FC switch.
• Fabric Provided Mac Address (FPMA) is a 48-bit mac-address that uniquely identifies a single VN_port(CNA) on an Enode. FPMA is assigned to an VN_port by FC switch during FLOGI process.
• FPMA (48-bit mac) = FC-map (24 bits) + FC_ID (24 bits)
  • FC-MAP: 24 bit value that identifies the FC switch. Default value: 0efc00
  • FC_ID:  Also called as N_port ID. A 24-bit value assigned to a VN_Port by the FC switch to uniquely identify it in the local fabric network.
• So, there will be one-one mapping between physical N-port WWPN and FPMA/N_port ID in a fabric network.

Configuring Qlogic CNA for FCoE on an ESXi Host

This article is about how to configure FCoE functionality on Qlogic CNA card-QL8262 installed on a Dell PowerEdge R720 server running ESXi 5.5.

QL8262 is dual port CNA card that supports LAN, iSCSI and FCoE. Configuring Qlogic FCoE card on an ESXi host is a two-step process;

1. Install driver on ESXi so that it can detect Qlogic cards.
2. Enable FCoE functionality on Qlogic CNA.